Week 12 Problem Session: Capstone review

Working through the Kuhn poker solve by hand at a small scale, then a closing discussion of what the semester's tools did and didn't capture

10 May 2027

Before the session#

Read the lecture page. Bring the −1/18 published Kuhn poker game value and the α-parameterized strategy table — the session traces two or three iterations of regret minimization by hand on one information set, to see concretely what “shift probability toward whatever was regretted most” actually does to the numbers.

In the session#

Working in pairs, the group picks one of the twelve Kuhn poker information sets and manually computes the regret-matched strategy for a few rounds from an arbitrary starting point, watching it drift toward the published equilibrium’s shape without ever being told the target values in advance.

Second half: the semester’s own scorecard, revisited. Each pair defends, using only what’s on the lecture page, whether they’d upgrade or downgrade this course’s own self-assessment of any of the four resolved claims — arguing with the material rather than just restating it.

Afterwards#

This is the last Problem Session of the semester. Nothing further is assigned here — the final synthesis this week’s lecture ends on is the last new idea; the remaining problem sets, the midterm, and the final exam are where it gets tested.